Maintaining a separate specialty vector database platform in addition to your existing database for structured data with PostgreSQL introduces significant cost, overhead, and complexity in managing multiple database systems, vendors, skill sets, and tools. 

Purpose-built vector databases may provide basic functionality for storing and retrieving vectors, but often lack robust data management features that are critical for production workloads: high performance indexing and querying, enterprise scalability, data integrity through ACID, disaster recovery, security, governance, regulator compliance, and more.

Vector data existing in one database and structured data in another creates data silos. Organizations then face complex data pipelines involving ETL processes or streaming integrations and joining data across these disparate systems for analysis.

EDB Postgres AI adds a "vector" data type to PostgreSQL, allowing vector embeddings from machine learning and AI models to be stored directly in the database as first-class data types.

With EDB Postgres AI, users can query and analyze vector data using standard SQL SELECT statements along with vector-specific operators and functions provided by the extension.

Vector data stored in EDB Postgres AI can be indexed using the database's indexing capabilities, enabling high-performance similarity searches and queries on vector embeddings.

EDB delivers an integrated data platform by providing vector database capabilities through EDB Postgres AI within the robust PostgreSQL database, allowing customers to leverage mature enterprise data services such as high availability and backup/recovery from a single vendor.